ThoughtSpot is putting a feedback loop in its AI tool ThoughtSpot Sage, which should make the tool’s analytics answers more accurate.
ThoughtSpot wants to improve Sage’s accuracy. Sage is the BI specialist’s generative AI tool. ThoughtSpot notes that generative AI is very good at interpreting natural language, but not necessarily at converting it into complex analytics queries. To do this, the AI must combine the right data sources in the right way, which exceeds the capabilities of classic LLMs. That’s why ThoughtSpot puts a feedback loop in its solution.
Specifically, users can view the AI’s answers in detail and adjust or improve queries. By adjusting prompts and providing better associations between terms and data, the model is immediately trained better. Sage helps with analytics questions and, in complex cases, provides an impetus towards an answer, after which data specialists can build on it.
ThoughtSpot also extends Sage with some other features. For example, Ask Sage makes it possible to ask follow-up questions based on visualizations, again in natural language. Sage can also monitor KPIs and notify users of important changes in natural language.
ThoughtSpot introduced Sage in March of last year. Now the company is building on the AI technology with new features and improvements.